基于几何不变矩的交通标志识别

以警告交通标志图像为研究对象,首先采用Hu不变矩来提取交通标志图像特征,然后利用最小矩距离法和相关系数法对交通标志进行识别。实验结果表明,该方法具有很好的识别能力,对旋转、缩放后的图象可以达到较高的识别率,基于Hu不变矩的几何特征提取算法在图像识别中的应用是非常有效的。     

从管道检测的角度,分析涂层失效的原因

本文首先阐明涂层对管道保护的重要作用，综述了目前长输管道涂层现状及常用涂层材料的特点，通过近几年来对数千公里管道的不同类型涂层检测、评价，总结出涂层失效的主要形式，从管道周围介质的腐蚀性、环境及涂层施工等方面，分析了外界条件与涂层失效的关系，以及导致涂层失效的主要原因。目前长输管道涂层失效的形式主要有：涂层的剥离、龟裂、破损及变形等。     

CSDF:一种基于STL的无锯齿差分网格剖分算法

针对基于差分的铸造CAE系统生成体网格时图像显示呈锯齿状这一问题,提出了一种基于STL面片模型的无锯齿网格剖分算法。算法采用一种新的组件式有符号距离场(CSDF)模型来对多个STL实体进行剖分。该模型采用了一种稳健的容错方法,能够对不封闭、缺点、少面的STL实体进行正确的剖分;使用BSP树来加速计算每个网格中心点距离铸件、砂芯、铸型以及浇注系统等STL几何实体的最小距离,并规定位于内部的中心点的距离为负,外部为正,表面为零,使得每一个网格点有一个表示材质的数据,同时还有几个不同的距离值;材质数据用来表示物理属性,而距离值用来进行STL实体的布尔运算,以实现组件式的网格剖分和几何模型的重建,获得光滑的后处理图像。实例表明:算法可行,可实现组件式的几何实体无锯齿网格剖分,且剖分结果能进行数值计算。     

一种基于清晰度的显微浮雕轮廓术

提出了一种基于清晰度的显微浮雕轮廓术。 采用5倍显微物镜,将正弦条纹通过投影仪从显 微镜目镜瞳口处耦合投射到载物台上,按照一个合适的步长移动Z轴,用CCD采集一系列不同 离焦量下 的条纹图像,用variance清晰度评价函数计算清晰度与离焦量之间的归一化曲线图,由此建 立清晰度与高 度的查找表;然后,放入待测浮雕物体,用CCD采集对应的条纹图像,采用图像分割分解不 同浮雕高度的 区域,再计算各区域内条纹的清晰度,根据查找表查找相应的高度,从而得到待测浮雕物体 不同区域的高 度信息,以此来重构浮雕物体的三维形貌。实验仿真分析证明了该方法的可行性,硬币上数 字为“0”的浮雕 物体表面三维形貌重建结果证明了该方法的有效性,其误差范围在[－2μm,2μm]之间 。     

应用于雷达水位计的等效采样扩时测距方法

雷达水位计是一种非接触式的水位测量仪器,以电磁波为载波方式测量传感器到水面的距离,再通过计算得出实际的水位值。由于电磁波传播速度接近光速,近距离传输时相应的时间间隔很短,要达到cm 级的距离测量精度,时间测量的精度需要达到ps。常用的高精度时间测量方法存在价格昂贵或实现方式复杂等问题。文中介绍了一种等效采样扩时测距方式,实现在雷达水位计的测量范围内达到cm 级测量精度,并已在产品中实现了具体应用。     

Anchor‐free distance estimation: A new approach to distance estimation for multihop ad hoc wireless networks

In ad hoc wireless networks, devices that normally cannot directly communicate route their messages through intermediate nodes. The number of those nodes is called hop count, a useful metric in estimating the distance between 2 nodes. Current methods usually depend on special nodes, called anchors, that need accurate localization information, in order to calculate an estimate for the average distance traversed per hop. The drawback of this approach is that anchor nodes increase the overall cost and complexity of the system. To address this problem, this letter proposes a novel, anchor node–free algorithm that can achieve a useful estimate for actual distance between 2 nodes, by analytically finding an estimate for the average maximum distance traveled per hop and multiplying with the hop count. The only requirement is the a priori knowledge of the networks' node density and the node range. The performance of our method is compared with a recent anchor node–based method and is shown to yield similar location estimation accuracy, despite the fact that it does not use anchor nodes.     

Distance and clustering‐based energy‐efficient pseudonyms changing strategy over road network

To improve the fairness, the energy consumption changing pseudonyms needs to be taken into account. Existing works focus on changing velocity‐based pseudonyms changing strategy and short changes interval with limited coverage, but due to similar velocity and short changes, internal attacker guesses easily known communication and location information due to location information of vehicle on tracking, which may expose adversary private information, and frequently, pseudonyms changing occurs due to movement of vehicles' similar velocity and short coverage, which may cause serious attack of vehicle. To overcome this problem, distance and cluster can be performed. In this work, we proposed distance and cluster‐based energy pseudonyms changing method for road network. We proposed distance and energy‐based clustering routing service over road network, the cluster head elected to depend on random number of distance and energy to change pseudonyms of vehicles. An each interval to be establish cluster head vehicle deployed while selects the operation mode and informs the cluster members of the selected mode through beacon signal. The cluster head vehicle node performs the pseudonyms changing based on the predicted distance and energy of the cluster member to use clustering optimization. The data of whole network send to report server through these nodes while near the RSU, and the vehicles in this area will use less energy to change the pseudonyms. The simulation results show that the proposed method enhances pseudonyms changing strategy less consumption and delays sufficient privacy level each vehicle also our method has outperform compare with existing methods than we use Sumo simulation and Matlab tools to verify our proposed method. Our proposed method outperformed in terms of pseudonym changing energy efficiency to careful attention during the cluster formation process, stable and balanced clusters that prolong the network lifetime, increases distances to more CH vehicles connectivity to makes clustering group and changing their pseudonyms in terms of high level privacy and finally, CH nodes use Dijkstra's algorithm use MST among the vehicles nodes depend on existing road networks to follow shortest path selection roads in terms of high connectivity probability of CH and stable structure of the network decreases the topology changes and thus,the clustering overhead is reduced.     

A method for throughput enhancement in Wi-Fi backscatter communications using power-level modulation

An improved data rate technology from tag to reader in Wi-Fi backscatter systems is proposed. The proposed system applies a power-level modulation to the conventional backscatter communication system. This technology can improve the data rate dramatically. It also adopts the Euclidean distance evaluation during demodulation, which looks for the most similar signal to enhance the BER performance. We analyze the error performance of the proposed decoding method, and the numerical results are presented. The power level is formed ranging from 2-level to 8-level. The performance of the detecting algorithm is compared to the conventional system, which utilizes the received power of selected sub-carriers.     

基于双向LSTM的误植域名滥用检测方法

当前,误植域名检测主要以计算域名对之间的编辑距离为基础,未能充分挖掘域名的上下文信息,且对短域名的检测易产生大量的假阳性结果。采集域名相关信息进行判定虽然有助于提高检测效果,却会引入较大的额外开销.本文采用了基于域名字符串的轻量级检测策略,并引入双向长短时记忆模型（LSTM,Long Short-Term Memory）来充分利用域名上下文,提升检测效果.本文还设计了面向域名的局部敏感哈希函数,以提高在大规模域名集合上进行误植域名检测的速度.在大量真实数据集上的实验结果表明,本文的工作改进了基于编辑距离检测方法的不足,能够有效地进行误植域名滥用检测.     

AdaCode:IEEE 802.11n长距离无线链路自适应编码算法

通过在实验床上的大量真实测量,分析了长距离无线链路中导致吞吐率低下的原因,总结了数据帧出错模式和丢包特点.在此基础上,提出了MAC（Media Access Control）层自适应编码算法AdaCode.该算法将帧间编码和帧内编码有机结合起来,帧间编码使用喷泉码LT（Luby Transform）码,帧内编码使用RS（Reed Solomon）码,根据链路中数据帧误比特率和丢包率的实时变化,动态选择最佳编码率,使接收端能够解码出所有数据帧,从而避免重传,极大提升了链路吞吐率.在MATLAB平台上实现了AdaCode,并使用实验床测得的真实数据进行仿真.仿真结果表明在固定物理层发送速率和可变发送速率两种情况下,AdaCode都能有效提升链路吞吐率.